Comparative study of classification method on diagnosis of plasmodium phase

Abstract

Malaria is a plague in humans induced by the Plasmodium parasitoid transferred through a single bite of female Anopheles mosquitoes. Once reaching human blood passage, this parasitoid bears asexual proliferation which is classified toward three phases (trophozoite, schizont, and gametocyte). To discover the phases, the paramedic will analyse the blood specimen of the subject within the microscope. However, the aforementioned approach has the potential to misdiagnosis. Many CAD-based investigations have been carried out to reduce the diagnosis errors that have formed. This research sought to develop a CAD-based design that can help paramedics in diagnosing the parasitoid Plasmodium, including trying to get a classification algorithm that is able to analyze the Plasmodium phase precisely. Based on the experiment outcome, Naïve Bayes was mostly useful applied as a classifier that attained the accuracy of 97.29%, the sensitivity, and specificity of 97.30% toward the P.vivax case. In the case of P.falciparum, it emerged in the accuracy, sensitivity, and specificity of 98.36%, 98.40%, and 98.40% apiece. Meantime, Perceptron was a useless algorithm applied as a classifier that realized the accuracy up to 81.08%, sensitivity and specificity of 93.80% on each. In the case of P.vivax, the accuracy, sensitivity, and specificity realized were 80.33%, 92.50%, and 92.50% respectively.

Comparative Analysis of SVM and Perceptron Algorithms in Classification of Work Programs

“…It can also be used for the perceptron algorithm with the sigmoid activation function in performing comparative analysis. In reference to previous research, the accuracy of SVM was obtained with a value of 98.28% from [9], with a value of 88.2% [10], with a value of 92.3% from [11], then reference to previous research obtained a perceptron accuracy value with a value of 81.1% from [12], with a value of 80.5% from [13], with a value of 85.3% from [14]. Data mining is one of the sciences in computer science, which can be in the form of grouping to be used in solving problems such as predictions, classification, or regression [15,16].…”